CHANGING FUEL FORMULATIONS WILL BOOST HYDROGEN DEMAND

Knut A. Simonsen, Luke F. O'KeefeTexaco Inc. White Plains, N.Y.
W. Francis FongTexaco Development Corp. White Plains, N.Y.

Refinery demand in the U.S. for on-purpose hydrogen will continue to increase by 5-10%/year, depending on the extent of implementation of the 1990 U.S. Clean Air Act Amendments (CAAA) and other proposed environmental legislation.

The two processes capable of meeting this demand are steam methane reforming (SMR) and partial oxidation.

SMR is essentially a mature process, with only incremental improvements in efficiency and reliability expected. Emissions from burning fuel to supply heat for this endothermic process and limited feedstock flexibility continue to be concerns, especially if the price of natural gas in the U.S. rises over the next decade.

CLEAN AIR ACT

U.S. refiners are carefully examining the impact the CAAA will have on their operations. Although the debate on the economic wisdom of the legislation still rages, it is evident that refiners likely will see a large upswing in hydrogen demand while existing hydrogen production may decline.

To better understand the potential impact various reformulation scenarios may have on the refining industry, and specifically, on the demand for hydrogen, Texaco analyzed the hydrogen supply/demand scenario in great detail.

It is critical for the competitiveness of the U.S. economy that the public and the federal and state legislators understand which changes to gasoline and diesel fuels will be most cost-effective. The optimal solution will include a combination of reformulated fuels, incentives for retirement of older cars, effective inspection and maintenance laws, and cost-effective changes to automobile hardware (vapor-recovery canister, preheated catalyst, etc.).

The aim of any program should be to obtain the greatest benefits for the cost, and to do it by finding the best balance among energy, economics, and the environment.

Two cases were studied in this analysis: mild and severe reformulation. The mild reformulation case is based on current CAAA legislation along with minor modifications to automobile hardware. The severe case is based on a nationwide implementation of Phase 2 of the CAAA and California's proposed reformulated fuels.

HYDROGEN DEMAND

The required changes in gasoline and diesel specifications could lead to a sharp increase in the demand for on-purpose or dedicated hydrogen production, and could end up costing the U.S. refining industry more than $3 billion in capital investments.

Refiners may have to reduce total aromatics, heavy ends (90% distillation point, or T90), and vapor pressure in gasoline by the end of the decade. These requirements will lead to large increases in the use of oxygenates-such as methyl tertiary butyl ether (MTBE)-and hydrogen for upgrading gasoline.

Diesel fuels are facing similar pressures. The push to reduce sulfur and aromatics, along with a requirement to increase the cetane level, will lead to a significant increase in hydrogen demand.

As noted earlier, refiners have two commercially viable options for producing additional manufactured hydrogen: steam methane reforming and partial oxidation. While both technologies provide high-purity hydrogen, partial oxidation can do so with reduced capital cost, feedstock flexibility, and without the emissions problems (such as Nox) normally associated with SMR.

Current estimates put the U.S. hydrogen market at approximately 3,670 bscf/year.1 Of this demand, the refining industry is the largest end user, consuming about 60% of the supply. Fig. 1 shows hydrogen-consuming industries and their respective demands.

CURRENT BALANCE

Five states comprise two thirds of U.S. crude oil refining capacity: Texas, Louisiana, California, Illinois, and Pennsylvania. Among these five states with the largest capacities, California has the largest on-purpose hydrogen production capacity (Fig. 2).

California accounts for 43% of on-purpose hydrogen capacity and Texas accounts for 25% (2.5 bscfd). Taking into consideration estimated by-product hydrogen production from the catalytic reforming unit (CRU), the total hydrogen produced in California accounts for only 25% of U.S. capacity (6.0 bscfd), while Texas accounts for 28%.

California still produces 60% more hydrogen (on-purpose plus by-product) than Texas, per barrel of crude refined. This is attributed to the heavier crudes refined in California and the more stringent environmental standards, which may soon impact the rest of the nation.

U.S. supply and demand for hydrogen are currently close to being in balance. More than 10 out of the 184 active U.S. refineries, however, are short more than 15 MMscfd.

It is possible to derive an approximate hydrogen balance for individual refineries using industry-established data.2

Hydrogen-consuming processes in the refinery have the following typical hydrogen requirements:

Catalytic hydrocracking-1,000-2,500 scf/bbl of feed
Catalytic hydrorefining-100-1,000 scf/bbl
Catalytic hydrotreating-50-600 scf/bbl.

CRUs typically yield 1,000-1,500 scf by-product hydrogen/bbl of feed.

By repeating the hydrogen balance calculation for all refineries (based on publicly available information), a detailed breakdown of the U.S. refinery hydrogen situation can be obtained.3

Table 1 presents the composite results of this analysis. All hydrogen supply and demand calculations are, for comparison, based on existing plant capacity rather than on an assumed availability.

By-product hydrogen recovered from the CRU can be lost as a result of being dissolved in liquids, lost in vent gases, lost in purification steps, or attributed to a reserve margin. After factoring in these losses, some refineries have significant shortages of hydrogen.

Less than 30 out of the 184 active U.S. refineries have significant excess hydrogen supply (15 MMscfd). In the Gulf Coast region, some refineries will be able to purchase smaller quantities of hydrogen from existing hydrogen pipelines and from the ethylene industry (OGJ, Aug. 24, 1992, p. 26).

DEMAND GROWTH

Total U.S. refinery hydrogen demand is currently 6.0 bscfd, 60% of which is recovered in the refinery (mostly as a by-product from the CRU). The remaining 40% is recovered from on-purpose hydrogen production facilities. This split is expected to be reversed in the next 5-10 years.

Historical demand for on-purpose hydrogen has increased at 3.5%/year over the past 10 years (Fig. 3). Demand for on-purpose hydrogen is projected to grow from 2.5 bscfd in 1991 to 3.7-6.2 bscfd over the next 5-10 years, depending on the composition of federal and regional regulations (Fig. 4). These increases are equivalent to a 5-10%/year increase.

The largest increase in demand for on-purpose hydrogen occurs in the severely reformulated case. Taking such a course would lead to a drop in by-product hydrogen supply from the CRU because of the proposed reductions in gasoline aromatics. And total hydrogen demand would increase as well, as a result of the changes to gasoline and diesel formulations (Fig. 5). Total hydrogen production capacity is projected to grow from 6.0 bscfd in 1991 to 6.9-8.4 bscfd over the next 510 years, to keep up with demand (Fig. 5).

The mild reformulation case yields a more cost-effective solution to dealing with automobile emissions than does the severely reformulated case. The mild case also leads to an additional demand for 1.2 bscfd of on-purpose hydrogen, while demand in the severe case increases by 3.7 bscfd.

Depending on the strategy chosen, 40-120 new hydrogen facilities (average size, 30 MMscfd) could be needed within this decade. The additional need for hydrogen could cost the U.S. refining industry, and eventually the consumer, $1-4 billion.

REFORMULATED GASOLINE

In 1995, the CAAA targets nine metropolitan areas in the U.S. with relatively high levels of ozone in the atmosphere: Los Angeles, San Diego, New York, Philadelphia, Hartford, Baltimore, Houston, Chicago, and Milwaukee. These areas comprise roughly one third of the U.S. gasoline pool.

This regulation will probably lead to some additional hydrotreating of catalytic naphtha and a scale-back on reformer severity on a regional basis.

Some groups are advocating nationwide reformulation of gasoline without evaluating the various alternatives on a cost/benefit basis. If severe reformulation were implemented nationwide, the entire 7.3 million b/d gasoline pool would be affected.

Severe reformulation could require total aromatics content to be decreased from 32 to 25%, along with a possible reduction in heavy ends (T90). And large increases in the use of oxygenates could result in adding 11-15 vol % to the gasoline pool.

If implemented nationwide, these changes could require almost 1.0 million b/d of oxygenates, effectively replacing a large portion of octane-boosting components such as aromatics and heavy ends. These components are produced primarily by the fluid catalytic cracking unit (FCCU) and the catalytic reforming unit.

The reduction of aromatics, accomplished by a decrease in reformer severity, could result in an equivalent 20-30% drop in by-product hydrogen production. This is equivalent to 0.85-1.25 bscfd of hydrogen that would have to be replaced by on-purpose production. Consequently, the increase in heavy ends from the reformer may have to be hydrotreated, which would require additional hydrogen.

The nationwide reformulation of gasoline would also lead to the reduction of gasoline Rvp from approximately 9.6 psi to less than 8.5 psi-even lower in summer months. The result may be a large surplus of heavy ends from the FCCU.

Potential reductions in T90 after 1995 could dramatically reduce the quantity of FCCU naphtha available for the gasoline pool. Reducing T90 would be very expensive and would likely have little impact on improving air emissions.

FCCU naphtha, at about 35 vol %, is by volume, the largest gasoline component. The FCCU is also, along with the reformer, a large contributor of aromatics.

If the 90% point for gasoline is reduced from about 350 F. to 300 F., then the 10 vol % of heavy-cut catalytic naphtha would have to be hydrotreated or hydrocracked, producing diesel or possibly the higher value-added jet fuel.4 This would require at least 1,500 scf of hydrogen/bbl feed.

A 10 vol % reduction in naphtha would produce about 550,000 b/d of heavy catalytic naphtha. To saturate this aromatic feed, approximately 0.80 bscfd of additional hydrogen would be required.

Another option for refiners would be to send the aromatics stream to a partial oxidation hydrogen-generation unit.

FUTURE DIESEL SPECS

Current legislation calls for the aromatics content in California diesel to be reduced from about 33% to 10% by October 1993. There is, however, an alternative engine certification procedure that may, along with a cetane improver, allow higher aromatic content fuel (25-30%).

Smaller refineries (

The 1990 CAAA requires diesel sulfur levels to be reduced nationwide from about 0.35 wt % to 0.05 wt % by October 1993. This reduction will require less than 0.15 bscfd hydrogen.

In contrast to gasoline engines, diesel engines require fuels with a low self-ignition temperature. This makes paraffins, with their low self-ignition temperature, more desirable, and aromatics less attractive.

Diesel's cetane number is a quality that parallels gasoline's octane number. The CAAA requires a minimum cetane index of 40, corresponding to a cetane number of about 45.

Again, the need for lighter products and a higher cetane number will increase hydrogen demand. Assuming a few other states adopt the California rules, the additional need for hydrogen for diesel reformulation would be in excess of 0.6 bscfd. If the rest of the country were to follow the California diesel laws, there would be an even larger increase in demand for on-purpose hydrogen.

Reducing aromatics to 10 vol % would require approximately 1,000 scf additional hydrogen/bbl diesel.5 Based on a total diesel market of approximately 1.5 million b/d, 1.5 bscfd additional hydrogen would be needed. Conversely, if aromatics content were reduced to 20 vol %, then about 0.9 bscfd of hydrogen would be needed.

Table 2 presents an overview of the increase in demand for on-purpose hydrogen in relation to potential changes in gasoline and diesel fuel composition.

OTHER IMPACTS

This hydrogen outlook does not take into account these issues, which may further add to hydrogen demand:

The U.S. crude slate is expected to get heavier and higher in sulfur (OGJ, Mar. 1, p. 62), while the demand for lighter products is increasing.
This will require additional hydrogen to saturate the heavy end of the barrel.
Many existing on-purpose hydrogen facilities (mostly steam methane reformers) are becoming outmoded, and in some cases it will be more economical to replace these facilities with new capacity.
Foreign export-oriented refineries will also need to improve the quality of the gasoline and diesel they now sell into the U.S. market, increasing offshore demand for hydrogen.

HYDROGEN PRODUCTION

While there are a variety of methods of manufacturing hydrogen, steam methane reforming and partial oxidation are best suited for large-volume refinery applications.

Other methods are either too expensive (often exceeding $5/MMBTU) or still at an experimental stage. These include water electrolysis, the internal-combustion engine process (oxidation of methane by oxygen), water splitting (thermal splitting, thermochemical cycles, steam-iron cycles, and nuclear radiolysis), and experimental solar processes.

STEAM METHANE REFORMING

Steam methane reforming of hydrocarbons-primarily natural gas-continues to be studied extensively and has been employed widely for on-purpose hydrogen generation in refineries. The reactions typically chosen to represent the steam reforming of methane, the primary constituent of natural gas, are:

CH4 + H2O CO + 3H2 (1)

CO + H2O CO2 + H2 (2)

Overall, SMR is an endothermic reaction, with the methane/steam reforming reaction (1) favored by high temperatures and low pressures (

SMR technology is mature, with numerous technology vendors competing on the basis of reformer furnace design (top, side, or bottom-fired), steam-to-carbon ratio, CO conversion and hydrogen purification systems, piping design and metallurgy, and other proven design concepts.

Reforming technology is limited in practice to natural gas or light naphtha feedstocks, as the presence of catalyst poisons (sulfur, chlorides, etc.) and heavier olefinic feeds must be strictly avoided to stop coking problems within the reformer tubes.

The relatively recent addition of a pressure swing adsorption (PSA) unit to the steam reforming process was a significant step in simplifying the hydrogen purification step (Fig. 6).

Development work continues on the reforming process, primarily focused on heat recovery, such as:

Prereformer catalyst vessel to improve feedstock flexibility
Gas-heated reformer (GHR) developed by ICI for optimal heat flux
Secondary reformer using oxygen for increased efficiency.

The primary disadvantages of the steam reforming process are the firing of a portion of the natural gas to provide the heat of reaction for this endothermic process, and the emissions this firing generates. While the addition of a selective catalytic reduction (SCR) unit to clean up the flue gas can reduce NOx emissions, it may not be sufficient to meet current and proposed limits in many states, including California (OGJ, Nov. 2, 1992, p. 45).

The restriction on product pressure (

PARTIAL OXIDATION

Partial oxidation (gasification) technology was first developed and commercialized by Texaco in the 1940s. The process uses natural gas and pure oxygen to produce a mixture of hydrogen and carbon monoxide, commonly called synthesis gas or "syngas."

Since the '40s, this technology has been further developed to process refinery offgases, heavy vacuum residue, petroleum coke, and even coal, as feedstocks.

Today many refineries are taking a second look at partial oxidation technology for their hydrogen needs because:

Partial oxidation uses no furnace and therefore produces no NOx emissions.
Partial oxidation can accept a wide variety of feedstocks, including sulfur-containing refinery offgas, cracked gas, FCCU offgas, and other waste streams. If a low-value (liquid or solid) feedstock such as vacuum resid or coke is available, co-production of hydrogen, steam, and electricity may be an attractive alternative.
U.S. refineries will be long on purge gases, butanes/propanes, T90-cut liquids, and heavy by-product streams as a result of changing environmental laws. This will necessitate some method of using these streams.
More refiners are planning to use oxygen instead of air for Claus unit expansions and FCCU upgrades. Furthermore, the use of nitrogen instead of natural gas for equipment purging or blanketing at unit shutdown and start-up is increasing. As refinery demand for oxygen and nitrogen increases, the need for an on site air separation unit will become important.

The partial oxidation reaction is exothermic and is carried out very rapidly at high temperature (2,200-2,400 F.):

CxHy + x/2O2

xCO + y/2H2 (exothermic)

Besides this reaction, some minor side reactions also occur at specific operating conditions:

CO + 1/2O2 CO2

H2 + 1/2O2 H2O

H2O + CO CO2 + H2

CH4 + 1/2O2 CO + 2H2

For a natural gas feed, syngas produced by partial oxidation contains mainly H2 and CO, with 2-3% CO2, 0.1-0.5% CH4, and traces of soot (unconverted carbon). All sulfur compounds such as mercaptan (RSH) will be converted to H2S and traces of COS, both of which are easily removed by conventional sulfur-removal methods.

The pressure of the gasifier can be designed to match the feed-gas supply pressure (800 psig for example), so as to minimize or eliminate hydrogen product compression.

The latest development of Texaco's partial oxidation process, the HyTEX process (Fig. 7), can reduce hydrogen production costs significantly.6 This patented process employs a Texaco noncatalytic quench gasifier for syngas production, followed by a high-temperature (600-800 F.) CO-shift unit for conversion of CO to CO2 and H2.

The required steam for this exothermic reaction is internally generated from the quench gasifier. Therefore, no external steam is required. This reaction can be simply described by:

CO + H2O CO2 + H2

The hydrogen gas stream exiting the CO-shift unit can be cooled in a series of heat exchangers and steam generators to recover heat before entering a PSA unit for final hydrogen purification to 99.9% purity. The impurities-including H2S, CO2, CO, CH4, and N2-are rejected at low pressure, along with some H2 gas.

The heat content of this rejected gas stream is 90-130 BTU/scf, which is ideally suited for feed to a low-BTU boiler to produce 600-psig super-heated steam with minimal NOx emissions.

Alternatively, this rejected gas stream can be sent to an acid-gas removal unit for recovery of both sulfur and CO2 as by-products and the rest recycled back to the gasifier.

As on-purpose hydrogen production from processes such as partial oxidation and steam methane reforming overtakes by-product production from the catalytic reformer, existing hydrogen manufacturing technology is likely to be further refined, and newer technologies commercialized.

REFERENCES

1. Chemical Economics Handbook, SRI International, February 1990, p. 743.5000 1.

2. "Process Handbook 1991," Hydrocarbon Processing.

3. Thrash, Lou Ann, "Worldwide Refining Survey," OGJ, Dec. 23, 1991, P. 39.

4. Stokes, Gary M., Wear, Charles C., Suarez, Wilson, and Young, George W., "Reformulated gasoline will change FCCU operations and catalysts," OGJ, July 2, 1990, p. 58.

5. Haun, Anderson, Kauff, Miller, and Stoecker, "The Efficient Refinery, Hydrogen Management in the 1990s," presented at 1990 UOP Technology Conference, Des Plaines, Ill.

6. Fong, W. Francis, and Quintana, Manuel, "HyTEX: A Novel Process for Hydrogen Production," presented at 1991 National Petroleum Refiners Association annual meeting, Mar. 17-19, 1991, San Antonio.